This was my second Ludum Dare, and I very nearly gave up.
Before the theme was announced I knew that I wanted to do something with ‘proper’ game models. You know, going into Blender, making a character model, rigging it, making a run/walk cycle then exporting that into Three.JS and displaying it in the game. I also knew I wanted to play with some neato post-processing filters like Bokeh (depth of field) shaders and so forth too.
The problem was I didn’t have a clue how to do any of those things.
In my preparation for Ludum Dare I made some base code (published on Github) which gave me a basic rendering pipeline which gave me DOF, Bloom and Antialiasing passes so that with a bit of luck I just had to throw some geometry at it. I also had a play with bones in Blender and after a bit of fighting managed to get a sausage wobbling on my screen. I thought I had everything I needed.
So then the theme is announced and it’s the theme I was desperately hoping it wouldn’t be: Connected Worlds. Agh. Mind blank. Couldn’t think of a thing. So plan B then was just to jump into Blender and start making some random assets. I knew I was going to meet the theme by mixing very unlikely things. So I made a tripod…
… and then I made Slug Lord…
.. and then I made a model of the author Jane Austen.
Jane and the tripod had a run cycle and Slug Lord was flapping his tentacles around. But all this? This took me an entire day. I had never modelled a humanoid before, never animated a run cycle, never dealt with Inverse Kinematics. It was incredibly frustrating.
On the second day I knew I needed to work on the actual game. So, naturally, I spent 4 hours writing code to generate text in WebGL and load all the assets (I wanted to make sure all the assets had loaded before starting the game because, by default, Three.JS is very happy to start rendering things immediately and let textures and models pop in as they’re loaded).
Finally, with all that done and 14 hours remaining I needed to start on the game. I knew I wanted Slug Lord in the middle of the scene, with Jane running circles around him. The original vague idea was some sort of single screen point and click adventure, where the player character had to escape from the clutches of slug lord… there was no chance of that though. Instead I considered some sort of Canabalt style endless runner/jumping thing.. so I started putting the scene together… and naturally spent another 4 hours tweaking the depth of field effect which, for some reason, was refusing to work properly.
I approached breaking point here and wanted to just give up. I didn’t want to abandon the thing that had caused me to waste so much time, but I was, in fact, wasting huge amounts of time on a pointless effect most people wouldn’t even notice, and I saw the deadline approaching fast.
And I still didn’t even have a game.
Agh.
Mercifully though I got the effect working in the end. Just. At which point I was back in Blender to make a very rough and quick background scene.
At this point I had about 5 hours remaining… and still no game. The platform/jumping game was out because I didn’t have enough time, so I added my tripod model into the scene and realised I could probably just make a simple twitch reflex chase game. You know, press the button at the right time and you go faster, press the wrong button at the wrong time you go slower and then the tripod can catch you.
Simple. Kinda funny. Easy to implement. Messages popped up as instructions from Jane telling you which mouse button to click and the mechanics of the chase were in and… mostly balanced.
Had enough time, in fact, to get a neato shader in for the background and add a bunch of random particles and lights… And it was sort of done. And I got this:
And lo, it was silly. The red and green and snow thing was supposed to connect the world of Christmas to the worlds of The Slug Lord and Jane Austen but I guessed that was a bit too subtle. My work was done anyway.
Now, someone in the comments for this mentioned it probably should have been a typing of the dead type game. And this was the biggest ‘doh’ moment ever for me. Of course. Jane Austen was a writer. It would have been perfect. And if I’d thought of that on day one I could have done that. The time I spent tweaking shaders or fighting with weird vertex twisting in Blender could have been spent on a typing game mechanic and importing a load of Austenesque phrases and words. And that would have been great.
So lessons learnt: Never use technology you’re not 100% familiar or comfortable with. Ludum Dare is best when you’re spending the time coding up interesting game mechanics, not shader effects or learning new platforms or entire skills. Also: More base code, especially if you’re using Three.JS which puts you at a massive disadvantage compared with Unity.
Finally, and I made this same mistake last time and failed to learn from it : Make a lo-fi game with simple boxes and if there’s time add models and shiny later. Game first, graphics second. This is Ludum Dare, not EA.
Considering my original objectives I think I got what I wanted out of this Ludum Dare though. Disappointed the game itself is basically the thinnest possible layer on top of what is basically a screensaver, but considering what I had to learn to build this it’s possible – just possible – I got more out of it than I realise. Maybe.
The game’s page is here and typical play sessions last about 30 seconds. Which, you know, is fair enough.
Tags: postmortem ld30
(it’s take you 30 seconds to play it)
Fire beats Earth, Earth beats Water, Water beats Fire. Have fun! Btw, this is our second game for Ludum Dare.
